
# Import the python pandas library
import pandas as pd
import numpy as np
from numpy import nan as NA
import seaborn as sns
import matplotlib.pyplot as plt
import plotly.express as px
import nltk
from nltk.sentiment.vader import SentimentIntensityAnalyzer
nltk.download("vader_lexicon")
nltk.download('stopwords')
nltk.download('wordnet')
import imageio
from wordcloud import WordCloud
from wordcloud import STOPWORDS
from sklearn.feature_extraction.text import CountVectorizer, TfidfTransformer
stopwords = STOPWORDS
%matplotlib inline

[nltk_data] Downloading package vader_lexicon to
[nltk_data]     C:\Users\Aditya\AppData\Roaming\nltk_data...
[nltk_data]   Package vader_lexicon is already up-to-date!
[nltk_data] Downloading package stopwords to
[nltk_data]     C:\Users\Aditya\AppData\Roaming\nltk_data...
[nltk_data]   Package stopwords is already up-to-date!
[nltk_data] Downloading package wordnet to
[nltk_data]     C:\Users\Aditya\AppData\Roaming\nltk_data...
[nltk_data]   Package wordnet is already up-to-date!


# Load the items dataset
items_df = pd.read_csv("items.csv")
items_df.rename(columns = {"asin": "product_id"}, inplace = True)
items_df


# Load the reviews dataset
reviews_df = pd.read_csv("reviews.csv")
reviews_df.rename(columns = {"asin": "product_id"}, inplace = True)
reviews_df


# Merge the two dataframes
project_df = pd.merge(items_df, reviews_df, how = "right", on = "product_id")
project_df


# Rename the Columns
project_df.rename(columns = {"title_x": "product_title","rating_x": "product_rating", "rating_y": "review_rating", "title_y": "review_title"}, inplace = True)
project_df.head()


# Check for any duplicate rows in the dataframe
print(f"The total number of duplicated rows in the dataframe: {project_df.duplicated().sum()}")
project_df[project_df.duplicated()]

The total number of duplicated rows in the dataframe: 12


# Drop the duplicate rows
project_df.drop_duplicates(inplace = True)
project_df.head()


# Check for null values in each column in the dataframe
project_df.isnull().sum()

product_id            0
brand               200
product_title         0
url                   0
image                 0
product_rating        0
reviewUrl             0
totalReviews          0
price                 0
originalPrice         0
name                  2
review_rating         0
date                  0
verified              0
review_title         14
body                 21
helpfulVotes      40763
dtype: int64


# Delete the rows with null values in the columns: 'brand', 'name', 'review_title' and 'body'
project_df.dropna(subset = ['brand', 'name', 'review_title', 'body'], inplace = True)
project_df.drop(columns = ["helpfulVotes", "originalPrice"], axis = 1, inplace = True)
# Reset the index
project_df.reset_index(drop = True, inplace = True)
project_df


# Convert date column to a DateTime object
project_df["date"] = pd.to_datetime(project_df["date"])
project_df["review_year"] = project_df["date"].apply(lambda date: date.year)
project_df.head()


# Determine the number of cellphones for each brand
items_df.groupby(by = "brand")["title"].count()

brand
ASUS          5
Apple        63
Google       38
HUAWEI       32
Motorola    105
Nokia        44
OnePlus      10
Samsung     346
Sony         27
Xiaomi       46
Name: title, dtype: int64


# Determine the average product ratings for each cellphone brand
items_df.groupby(by = "brand")["rating"].mean().sort_values(ascending = False)

brand
Xiaomi      4.415217
HUAWEI      4.021875
ASUS        3.860000
Sony        3.788889
Apple       3.782540
Google      3.771053
Motorola    3.643810
Samsung     3.632659
OnePlus     3.580000
Nokia       3.386364
Name: rating, dtype: float64


# Create a new price dataframe with correct prices
price_df = project_df[(project_df['price']!=0) & (project_df['price']!=1)]
# Plot the scatterplot for price vs product rating for each cellphone brand
fig = px.scatter(price_df, x="product_rating", y="price",  color='brand', title = "Price v/s Product_Rating")
fig.show()


# Determine the average product_rating for each brand and every year
data_rating_year=project_df.groupby(['brand', 'review_year'])[['product_rating']].mean().reset_index()
data_rating_year


# Plot the line graph for product_rating for each brand accross years
sns.set_style('darkgrid') # Set the grid style
sns.set(rc = {'figure.figsize':(20,10)}) # Set the figure size
sns.set_context('talk') # Set context
ratings_plot=sns.lineplot(data=data_rating_year, x='review_year', y='product_rating', hue='brand', style='brand', markers=True, dashes=False)
ratings_plot.set_title('Average Product Ratings by Year') # Set the title
ratings_plot

<AxesSubplot:title={'center':'Average Product Ratings by Year'}, xlabel='review_year', ylabel='product_rating'>


# Create the Sentiment Score columns for review title and body with the NLTK Vader Sentiment Analysis
analyzer = SentimentIntensityAnalyzer()
project_df['Title_Sentiment_Score'] = project_df['review_title'].apply(lambda x: analyzer.polarity_scores(x)['compound'])
project_df['Body_Sentiment_Score'] = project_df['body'].apply(lambda x: analyzer.polarity_scores(x)['compound'])
project_df.head(20)


# Creating a categorial variable as Positive, Negative or Neutral for reviews
project_df["Body_Sentiment_Type"] = np.where(project_df["Body_Sentiment_Score"] >= 0.05, "Positive", np.where(((project_df["Body_Sentiment_Score"] < 0.05) & (project_df["Body_Sentiment_Score"] > -0.05)), "Neutral", "Negative"))
project_df.head(15)


# Plot the Sentiment Types for each brand
plt.figure(figsize = (25,10))
sentiment_plot = sns.countplot(data = project_df, x = "brand", hue = "Body_Sentiment_Type")
plt.title("Sentiment Type for each Brand")
for p in sentiment_plot.patches:
    plt.annotate(f'{p.get_height()}', xy=(p.get_x() + p.get_width() / 2, p.get_height()), ha='center', va='bottom')


# Read the image
mask_image = imageio.imread('mask_oval.jpeg')
# Create a new data frame with only the reviews and the sentiment type for body
text_df = project_df[['body','Body_Sentiment_Type']].copy()
# Convert positive and negative reviews into long strings that can be analyzed
positive_df = text_df[text_df['Body_Sentiment_Type'] == 'Positive']
positive_text_list = positive_df['body'].values.tolist()
positive_text = ''.join(positive_text_list)
negative_df = text_df[text_df['Body_Sentiment_Type'] == 'Negative']
negative_text_list = negative_df['body'].values.tolist()
negative_text = ''.join(negative_text_list)


# Plot the postive review wordcloud
positive_wordcloud = WordCloud(colormap='prism', mask=mask_image, background_color='white', max_font_size=300, stopwords=stopwords)
positive_wordcloud = positive_wordcloud.generate(positive_text)
plt.figure(figsize=(20,10))
plt.title('Positive Review Word Cloud', fontsize = 32, fontweight = 'bold')
plt.grid(False)
plt.imshow(positive_wordcloud)

<matplotlib.image.AxesImage at 0x20487878a90>


# Plot the negative review wordcloud
negative_wordcloud = WordCloud(colormap='prism', mask=mask_image, background_color='white', max_font_size = 300, stopwords = stopwords)
negative_wordcloud = negative_wordcloud.generate(negative_text)
plt.figure(figsize=(20,10))
plt.title('Negative Review Word Cloud', fontsize = 32, fontweight = 'bold')
plt.grid(False)
plt.imshow(negative_wordcloud)

<matplotlib.image.AxesImage at 0x2049b6a25b0>


# Create a new dataframe with only product title, body and the Sentiment Type
phone_df = project_df[['brand','product_title','body','Body_Sentiment_Type']]
# Convert the positive and negative review text of Apple, Samsung, Motorola, and Xiaomi into seperate lists
apple_positive_list = phone_df[(phone_df['Body_Sentiment_Type']=='Positive') & (phone_df['brand']=='Apple')]['body'].values.tolist()
apple_negative_list = phone_df[(phone_df['Body_Sentiment_Type']=='Negative') & (phone_df['brand']=='Apple')]['body'].values.tolist()

samsung_positive_list = phone_df[(phone_df['Body_Sentiment_Type']=='Positive') & (phone_df['brand']=='Samsung')]['body'].values.tolist()
samsung_negative_list = phone_df[(phone_df['Body_Sentiment_Type']=='Negative') & (phone_df['brand']=='Samsung')]['body'].values.tolist()

motorola_positive_list = phone_df[(phone_df['Body_Sentiment_Type']=='Positive') & (phone_df['brand']=='Motorola')]['body'].values.tolist()
motorola_negative_list = phone_df[(phone_df['Body_Sentiment_Type']=='Negative') & (phone_df['brand']=='Motorola')]['body'].values.tolist()

xiaomi_positive_list = phone_df[(phone_df['Body_Sentiment_Type']=='Positive') & (phone_df['brand']=='Xiaomi')]['body'].values.tolist()
xiaomi_negative_list = phone_df[(phone_df['Body_Sentiment_Type']=='Negative') & (phone_df['brand']=='Xiaomi')]['body'].values.tolist()


# Store the CountVectorizer and the TfidfTransformer in variables
cv = CountVectorizer()
tfidf_transformer = TfidfTransformer()
# Create a list of positive and negative reviews for each of the four brands
positive_data_list = [apple_positive_list, samsung_positive_list, motorola_positive_list, xiaomi_positive_list]
negative_data_list = [apple_negative_list, samsung_negative_list, motorola_negative_list, xiaomi_negative_list]
# Define a list of the 4 brands
brand_list = ['Apple', 'Samsung', 'Motorola', 'Xiaomi']
positive_score = []
negative_score = []
# Calculate the TF-IDF score of the word 'battery' for the positive reviews of each brand
for phone_pos in positive_data_list:
  # Convert the list of positive review text into term-frequency matrix
  phone_data_pos = cv.fit_transform(phone_pos)
  # Convert term-frequency matrix into tf-idf matirx
  tfidf_matrix_pos = tfidf_transformer.fit_transform(phone_data_pos)
  # Create dictionary to find a tfidf word each word
  word2tfidf = dict(zip(cv.get_feature_names(), tfidf_transformer.idf_))
  for word, score in word2tfidf.items():
    if word in ('battery','Battery'):
      positive_score.append(score)

# Similarly, calculate the TF-IDF score of the word 'battery' for the negative reviews of each brand
for phone_neg in negative_data_list:
  phone_data_neg = cv.fit_transform(phone_neg)
  tfidf_matrix_neg = tfidf_transformer.fit_transform(phone_data_neg)
  word2tfidf = dict(zip(cv.get_feature_names(), tfidf_transformer.idf_))
  for word, score in word2tfidf.items():
    if word in ('battery','Battery'):
      negative_score.append(score)


# Make a plot of the TF-IDF value of the word 'battery' in both positive and negative reviews for top 4 brands
brand_name = ['Apple', 'Samsung', 'Motorola', 'Xiaomi']
width = 0.25
x = list(range(len(positive_score)))
plt.figure(figsize=(12,8))
# Create the bars for the TF-IDF scores of positive reviews
positive_plot = plt.bar(x, positive_score, tick_label = brand_name, label = 'Positive', width = width)
for i in range(len(x)):
  x[i] += width
# Create the bars for the TF-IDF scores of negative reviews
negative_plot = plt.bar(x, negative_score, tick_label = brand_name, label = 'Negative', width = width)
plt.xticks(size = 20)
plt.yticks(size = 20)
plt.xlabel('Phone Brands', fontsize = 24)
plt.ylabel('TF-IDF Scores', fontsize = 24)
plt.title('TF-IDF Scores for Positive and Negative Reviews on Battery for 4 Brands', fontsize = 24, fontweight = 'bold')
# Annotate the value for each bar
for p in positive_plot.patches:
    plt.annotate(f'{p.get_height():.2f}', xy=(p.get_x() + p.get_width() / 2, p.get_height()), ha='center', va='bottom', fontsize = 15)
for p in negative_plot.patches:
    plt.annotate(f'{p.get_height():.2f}', xy=(p.get_x() + p.get_width() / 2, p.get_height()), ha='center', va='bottom', fontsize = 15)    
plt.legend(fontsize = 15)
plt.show()

	product_id	brand	title	url	image	rating	reviewUrl	totalReviews	price	originalPrice
0	B0000SX2UC	NaN	Dual-Band / Tri-Mode Sprint PCS Phone w/ Voice...	https://www.amazon.com/Dual-Band-Tri-Mode-Acti...	https://m.media-amazon.com/images/I/2143EBQ210...	3.0	https://www.amazon.com/product-reviews/B0000SX2UC	14	0.00	0.00
1	B0009N5L7K	Motorola	Motorola I265 phone	https://www.amazon.com/Motorola-i265-I265-phon...	https://m.media-amazon.com/images/I/419WBAVDAR...	3.0	https://www.amazon.com/product-reviews/B0009N5L7K	7	49.95	0.00
2	B000SKTZ0S	Motorola	MOTOROLA C168i AT&T CINGULAR PREPAID GOPHONE C...	https://www.amazon.com/MOTOROLA-C168i-CINGULAR...	https://m.media-amazon.com/images/I/71b+q3ydkI...	2.7	https://www.amazon.com/product-reviews/B000SKTZ0S	22	99.99	0.00
3	B001AO4OUC	Motorola	Motorola i335 Cell Phone Boost Mobile	https://www.amazon.com/Motorola-i335-Phone-Boo...	https://m.media-amazon.com/images/I/710UO8gdT+...	3.3	https://www.amazon.com/product-reviews/B001AO4OUC	21	0.00	0.00
4	B001DCJAJG	Motorola	Motorola V365 no contract cellular phone AT&T	https://www.amazon.com/Motorola-V365-contract-...	https://m.media-amazon.com/images/I/61LYNCVrrK...	3.1	https://www.amazon.com/product-reviews/B001DCJAJG	12	149.99	0.00
...	...	...	...	...	...	...	...	...	...	...
715	B07ZPKZSSC	Apple	Apple iPhone 11 Pro, 64GB, Fully Unlocked - Sp...	https://www.amazon.com/Apple-iPhone-64GB-Fully...	https://m.media-amazon.com/images/I/41wDuEW9iZ...	1.0	https://www.amazon.com/product-reviews/B07ZPKZSSC	1	949.00	0.00
716	B07ZQSGP53	Xiaomi	Xiaomi Redmi Note 8, 32GB/3GB RAM 6.3" FHD+ Di...	https://www.amazon.com/Xiaomi-Display-Snapdrag...	https://m.media-amazon.com/images/I/41foh4FKHE...	4.6	https://www.amazon.com/product-reviews/B07ZQSGP53	3	150.96	0.00
717	B081H6STQQ	Sony	Sony Xperia 1 Unlocked Smartphone and WH1000XM...	https://www.amazon.com/Sony-Smartphone-WH1000X...	https://m.media-amazon.com/images/I/51zZTAXZTP...	4.5	https://www.amazon.com/product-reviews/B081H6STQQ	70	948.00	0.00
718	B081TJFVCJ	Apple	Apple iPhone X, 64GB, Gray - Fully Unlocked (R...	https://www.amazon.com/Apple-iPhone-64GB-Gray-...	https://m.media-amazon.com/images/I/71yMgOenT5...	5.0	https://www.amazon.com/product-reviews/B081TJFVCJ	1	478.97	0.00
719	B0825BB7SG	Samsung	Straight Talk Samsung Galaxy A10e Smartphone 5...	https://www.amazon.com/Straight-Samsung-Galaxy...	https://m.media-amazon.com/images/I/81+3SWSAhD...	5.0	https://www.amazon.com/product-reviews/B0825BB7SG	1	139.00	139.26

	product_id	name	rating	date	verified	title	body	helpfulVotes
0	B0000SX2UC	Janet	3	October 11, 2005	False	Def not best, but not worst	I had the Samsung A600 for awhile which is abs...	1.0
1	B0000SX2UC	Luke Wyatt	1	January 7, 2004	False	Text Messaging Doesn't Work	Due to a software issue between Nokia and Spri...	17.0
2	B0000SX2UC	Brooke	5	December 30, 2003	False	Love This Phone	This is a great, reliable phone. I also purcha...	5.0
3	B0000SX2UC	amy m. teague	3	March 18, 2004	False	Love the Phone, BUT...!	I love the phone and all, because I really did...	1.0
4	B0000SX2UC	tristazbimmer	4	August 28, 2005	False	Great phone service and options, lousy case!	The phone has been great for every purpose it ...	1.0
...	...	...	...	...	...	...	...	...
67981	B081H6STQQ	jande	5	August 16, 2019	False	Awesome Phone, but finger scanner is a big mis...	I love the camera on this phone. The screen is...	1.0
67982	B081H6STQQ	2cool4u	5	September 14, 2019	False	Simply Amazing!	I've been an Xperia user for several years and...	1.0
67983	B081H6STQQ	simon	5	July 14, 2019	False	great phon3, but many bugs need to fix. still ...	buy one more for my cousin	NaN
67984	B081TJFVCJ	Tobiasz Jedrysiak	5	December 24, 2019	True	Phone is like new	Product looks and works like new. Very much re...	NaN
67985	B0825BB7SG	Owen Gonzalez	5	December 11, 2019	False	Outstanding phone for the price	I love the size and style of this phone. It is...	NaN

	product_id	brand	title_x	url	image	rating_x	reviewUrl	totalReviews	price	originalPrice	name	rating_y	date	verified	title_y	body	helpfulVotes
0	B0000SX2UC	NaN	Dual-Band / Tri-Mode Sprint PCS Phone w/ Voice...	https://www.amazon.com/Dual-Band-Tri-Mode-Acti...	https://m.media-amazon.com/images/I/2143EBQ210...	3.0	https://www.amazon.com/product-reviews/B0000SX2UC	14	0.00	0.00	Janet	3	October 11, 2005	False	Def not best, but not worst	I had the Samsung A600 for awhile which is abs...	1.0
1	B0000SX2UC	NaN	Dual-Band / Tri-Mode Sprint PCS Phone w/ Voice...	https://www.amazon.com/Dual-Band-Tri-Mode-Acti...	https://m.media-amazon.com/images/I/2143EBQ210...	3.0	https://www.amazon.com/product-reviews/B0000SX2UC	14	0.00	0.00	Luke Wyatt	1	January 7, 2004	False	Text Messaging Doesn't Work	Due to a software issue between Nokia and Spri...	17.0
2	B0000SX2UC	NaN	Dual-Band / Tri-Mode Sprint PCS Phone w/ Voice...	https://www.amazon.com/Dual-Band-Tri-Mode-Acti...	https://m.media-amazon.com/images/I/2143EBQ210...	3.0	https://www.amazon.com/product-reviews/B0000SX2UC	14	0.00	0.00	Brooke	5	December 30, 2003	False	Love This Phone	This is a great, reliable phone. I also purcha...	5.0
3	B0000SX2UC	NaN	Dual-Band / Tri-Mode Sprint PCS Phone w/ Voice...	https://www.amazon.com/Dual-Band-Tri-Mode-Acti...	https://m.media-amazon.com/images/I/2143EBQ210...	3.0	https://www.amazon.com/product-reviews/B0000SX2UC	14	0.00	0.00	amy m. teague	3	March 18, 2004	False	Love the Phone, BUT...!	I love the phone and all, because I really did...	1.0
4	B0000SX2UC	NaN	Dual-Band / Tri-Mode Sprint PCS Phone w/ Voice...	https://www.amazon.com/Dual-Band-Tri-Mode-Acti...	https://m.media-amazon.com/images/I/2143EBQ210...	3.0	https://www.amazon.com/product-reviews/B0000SX2UC	14	0.00	0.00	tristazbimmer	4	August 28, 2005	False	Great phone service and options, lousy case!	The phone has been great for every purpose it ...	1.0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
67981	B081H6STQQ	Sony	Sony Xperia 1 Unlocked Smartphone and WH1000XM...	https://www.amazon.com/Sony-Smartphone-WH1000X...	https://m.media-amazon.com/images/I/51zZTAXZTP...	4.5	https://www.amazon.com/product-reviews/B081H6STQQ	70	948.00	0.00	jande	5	August 16, 2019	False	Awesome Phone, but finger scanner is a big mis...	I love the camera on this phone. The screen is...	1.0
67982	B081H6STQQ	Sony	Sony Xperia 1 Unlocked Smartphone and WH1000XM...	https://www.amazon.com/Sony-Smartphone-WH1000X...	https://m.media-amazon.com/images/I/51zZTAXZTP...	4.5	https://www.amazon.com/product-reviews/B081H6STQQ	70	948.00	0.00	2cool4u	5	September 14, 2019	False	Simply Amazing!	I've been an Xperia user for several years and...	1.0
67983	B081H6STQQ	Sony	Sony Xperia 1 Unlocked Smartphone and WH1000XM...	https://www.amazon.com/Sony-Smartphone-WH1000X...	https://m.media-amazon.com/images/I/51zZTAXZTP...	4.5	https://www.amazon.com/product-reviews/B081H6STQQ	70	948.00	0.00	simon	5	July 14, 2019	False	great phon3, but many bugs need to fix. still ...	buy one more for my cousin	NaN
67984	B081TJFVCJ	Apple	Apple iPhone X, 64GB, Gray - Fully Unlocked (R...	https://www.amazon.com/Apple-iPhone-64GB-Gray-...	https://m.media-amazon.com/images/I/71yMgOenT5...	5.0	https://www.amazon.com/product-reviews/B081TJFVCJ	1	478.97	0.00	Tobiasz Jedrysiak	5	December 24, 2019	True	Phone is like new	Product looks and works like new. Very much re...	NaN
67985	B0825BB7SG	Samsung	Straight Talk Samsung Galaxy A10e Smartphone 5...	https://www.amazon.com/Straight-Samsung-Galaxy...	https://m.media-amazon.com/images/I/81+3SWSAhD...	5.0	https://www.amazon.com/product-reviews/B0825BB7SG	1	139.00	139.26	Owen Gonzalez	5	December 11, 2019	False	Outstanding phone for the price	I love the size and style of this phone. It is...	NaN

	product_id	brand	product_title	url	image	product_rating	reviewUrl	totalReviews	name	review_rating	date	verified	review_title	body	helpfulVotes
0	B0000SX2UC	NaN	Dual-Band / Tri-Mode Sprint PCS Phone w/ Voice...	https://www.amazon.com/Dual-Band-Tri-Mode-Acti...	https://m.media-amazon.com/images/I/2143EBQ210...	3.0	https://www.amazon.com/product-reviews/B0000SX2UC	14	Janet	3	October 11, 2005	False	Def not best, but not worst	I had the Samsung A600 for awhile which is abs...	1.0
1	B0000SX2UC	NaN	Dual-Band / Tri-Mode Sprint PCS Phone w/ Voice...	https://www.amazon.com/Dual-Band-Tri-Mode-Acti...	https://m.media-amazon.com/images/I/2143EBQ210...	3.0	https://www.amazon.com/product-reviews/B0000SX2UC	14	Luke Wyatt	1	January 7, 2004	False	Text Messaging Doesn't Work	Due to a software issue between Nokia and Spri...	17.0
2	B0000SX2UC	NaN	Dual-Band / Tri-Mode Sprint PCS Phone w/ Voice...	https://www.amazon.com/Dual-Band-Tri-Mode-Acti...	https://m.media-amazon.com/images/I/2143EBQ210...	3.0	https://www.amazon.com/product-reviews/B0000SX2UC	14	Brooke	5	December 30, 2003	False	Love This Phone	This is a great, reliable phone. I also purcha...	5.0
3	B0000SX2UC	NaN	Dual-Band / Tri-Mode Sprint PCS Phone w/ Voice...	https://www.amazon.com/Dual-Band-Tri-Mode-Acti...	https://m.media-amazon.com/images/I/2143EBQ210...	3.0	https://www.amazon.com/product-reviews/B0000SX2UC	14	amy m. teague	3	March 18, 2004	False	Love the Phone, BUT...!	I love the phone and all, because I really did...	1.0
4	B0000SX2UC	NaN	Dual-Band / Tri-Mode Sprint PCS Phone w/ Voice...	https://www.amazon.com/Dual-Band-Tri-Mode-Acti...	https://m.media-amazon.com/images/I/2143EBQ210...	3.0	https://www.amazon.com/product-reviews/B0000SX2UC	14	tristazbimmer	4	August 28, 2005	False	Great phone service and options, lousy case!	The phone has been great for every purpose it ...	1.0

	product_id	brand	product_title	url	image	product_rating	reviewUrl	totalReviews	price	originalPrice	name	review_rating	date	verified	review_title	body	helpfulVotes
2527	B00836Y6B2	Nokia	Nokia Lumia 900 Black Factory Unlocked	https://www.amazon.com/Nokia-Lumia-900-Factory...	https://m.media-amazon.com/images/I/81ZlbLtZ3P...	3.2	https://www.amazon.com/product-reviews/B00836Y6B2	929	109.98	0.00	Dairon Alexander Maiz	5	February 13, 2015	True	exelente	exelente	NaN
3047	B00836Y6B2	Nokia	Nokia Lumia 900 Black Factory Unlocked	https://www.amazon.com/Nokia-Lumia-900-Factory...	https://m.media-amazon.com/images/I/81ZlbLtZ3P...	3.2	https://www.amazon.com/product-reviews/B00836Y6B2	929	109.98	0.00	Waygo	5	January 18, 2015	True	Five Stars	good	NaN
3073	B00836Y6B2	Nokia	Nokia Lumia 900 Black Factory Unlocked	https://www.amazon.com/Nokia-Lumia-900-Factory...	https://m.media-amazon.com/images/I/81ZlbLtZ3P...	3.2	https://www.amazon.com/product-reviews/B00836Y6B2	929	109.98	0.00	EDWIN MORENO	5	July 19, 2014	True	Five Stars	good	NaN
6985	B00E6FGSHY	Samsung	Samsung Galaxy S4, White Frost 16GB (AT&T)	https://www.amazon.com/Samsung-Galaxy-S4-White...	https://m.media-amazon.com/images/I/81suaO+v0m...	3.7	https://www.amazon.com/product-reviews/B00E6FGSHY	641	154.97	0.00	Northwest	5	February 26, 2015	True	Five Stars	Excellent phone. I love everything about it, s...	NaN
8378	B00F2SKPIM	Samsung	Samsung Galaxy Note 3, Black 32GB (Verizon Wir...	https://www.amazon.com/Samsung-Galaxy-Note-Ver...	https://m.media-amazon.com/images/I/91eFtaIWpc...	3.9	https://www.amazon.com/product-reviews/B00F2SKPIM	983	0.00	0.00	Duncan K.	5	March 6, 2014	True	Fantastic phone	i went from a s4 to the note 3 and all i can s...	1.0
14917	B00OZTSY6Y	Motorola	Motorola DROID Turbo XT1254, Black Ballistic N...	https://www.amazon.com/Motorola-DROID-Turbo-Ba...	https://m.media-amazon.com/images/I/81VMO0Upxf...	3.6	https://www.amazon.com/product-reviews/B00OZTSY6Y	545	149.99	0.00	B. KESLER	5	March 13, 2015	False	Best Phone on Verizon Right Now	I have long been a Motorola user, dating back ...	1.0
15628	B00V7FY44A	Samsung	Samsung Galaxy S6 Edge, White Pearl 32GB (Veri...	https://www.amazon.com/Samsung-Galaxy-S6-Edge-...	https://m.media-amazon.com/images/I/814WwhaRVu...	3.5	https://www.amazon.com/product-reviews/B00V7FY44A	505	0.00	0.00	Domingo Gomes - NewEsc	5	May 10, 2015	False	Five Stars	This is the best android Phone ever!!! I love ...	NaN
18163	B014GCG150	Samsung	Samsung Galaxy J5 SM-J500H/DS GSM Factory Unlo...	https://www.amazon.com/Samsung-SM-J500H-Unlock...	https://m.media-amazon.com/images/I/61alJun3Jv...	4.1	https://www.amazon.com/product-reviews/B014GCG150	645	148.98	0.00	Caleb Ajibade	5	July 8, 2019	True	Good	Good	NaN
27123	B01M0PADR4	Google	Google Pixel XL G2PW210032GBBK Factory Unlocke...	https://www.amazon.com/Google-G2PW210032GBBK-U...	https://m.media-amazon.com/images/I/71PZz7CQ9U...	3.4	https://www.amazon.com/product-reviews/B01M0PADR4	429	201.48	0.00	anon	5	January 10, 2017	True	Five Stars	Works great	2.0
33455	B06XYMCMHD	Samsung	Samsung Galaxy S8 SM-G950F Unlocked 64GB - Int...	https://www.amazon.com/Samsung-Galaxy-SM-G950F...	https://m.media-amazon.com/images/I/81FWIR3RbU...	3.9	https://www.amazon.com/product-reviews/B06XYMCMHD	117	399.99	0.00	Bulltwinkle	1	July 15, 2017	False	Amazon sells these phones without a US recogni...	Amazon sells these phones without a US recogni...	3.0
53176	B07HK4JNV1	Xiaomi	Xiaomi Redmi Note 6 Pro 64GB / 4GB RAM 6.26" D...	https://www.amazon.com/Xiaomi-Factory-Unlocked...	https://m.media-amazon.com/images/I/517Q3-wHBk...	4.3	https://www.amazon.com/product-reviews/B07HK4JNV1	441	179.50	0.00	vibhor jain	5	April 8, 2019	True	Nice phone	Awsome	NaN
66713	B07WSJYDXX	Motorola	Motorola G6 – 32 GB – Unlocked (AT&T/Sprint/T-...	https://www.amazon.com/Motorola-G6-Unlocked-T-...	https://m.media-amazon.com/images/I/51L6DbMbvK...	3.9	https://www.amazon.com/product-reviews/B07WSJYDXX	836	119.99	249.99	susan	4	July 15, 2019	True	Good.	Good.	NaN

Celltiment Analysis¶

Introduction¶

The project should be graded on the Data Analysis part because the datasets that I acquired didn't need that much cleaning and processing. Also, the depth of our analysis that I've done is more valuable as compared to the processing¶

Data Acquisition and Description:¶

Data Acquisition:¶

Data Description:¶

Data Processing:¶

Data Analysis¶

Sentiment Analysis¶

TF-IDF Analysis¶

Conclusions¶

	brand	review_year	product_rating
0	ASUS	2018	3.720468
1	ASUS	2019	3.945000
2	Apple	2015	2.800000
3	Apple	2016	3.180000
4	Apple	2017	3.878531
...	...	...	...
59	Sony	2017	3.655370
60	Sony	2018	3.728286
61	Sony	2019	3.864698
62	Xiaomi	2018	4.248000
63	Xiaomi	2019	4.393216